Unsupervised Learning for Inverse Problems in Computed Tomography

2026-03-19T09:48:19Z

Assume you encounter an inverse problem that shall be solved for a large number of data, but no ground-truth data is available. To emulate this encounter, in this study, we assume it is unknown how to solve the imaging problem of Computed Tomography (CT). An unsupervised deep learning approach is introduced, that leverages the inherent similarities between deep neural network training, deep image prior (DIP) and unrolled optimization schemes. We demonstrate the feasibility of reconstructing images from measurement data by pure network inference, without relying on ground-truth images in the training process or additional gradient steps for unseen samples. Our method is evaluated on the two-dimensional 2DeteCT dataset, showcasing superior performance in terms of mean squared error (MSE) and structural similarity index (SSIM) compared to traditional filtered backprojection (FBP) and maximum likelihood (ML) reconstruction techniques as well as similar performance compared to a supervised DL reconstruction. Additionally, our approach significantly reduces reconstruction time, making it a promising alternative for real-time medical imaging applications. Future work will focus on extending this methodology for adaptability of the projection geometry and other use-cases in medical imaging.

Mechanical cues for totipotency and the preneural state: embryo and cancer expanding the frontiers of developmental physics

2026-03-18T21:28:44Z

In this article, I advance the idea that physics plays a central role in cell differentiation and makes fundamental contributions to morphogenesis, revealing the totipotent nature of the zygote. Totipotency is a persistent mechanical memory that preserves the biomechanical records of animal morphogenesis. I examine the mechanical and biophysical pathways underlying cell differentiation in embryonic development and cancer, treating them as closely related biological and mechanical processes. Drawing inspiration from evolutionary history, I also propose a biophysical mechanism for the emergence of the animal nervous system. By linking physical principles to cellular differentiation, this review positions mechanobiology as a pillar of innovation with high-impact clinical implications for diseases such as cancer.

Extending the field of view in modulation-based X-ray phase microtomography

2026-03-18T16:48:56Z

Recent advances in propagation-based phase-contrast imaging, such as hierarchical imaging, have enabled the visualization of internal structures in large biological specimens and material samples. However, modulation-based techniques, which provide quantitative electron density information, face challenges when imaging larger objects due to stringent beam stability requirements and detector distortions. Extending the field of view of these methods is crucial for obtaining comparable quantitative results across beamlines and adapting to the smaller beam profiles of fourth-generation synchrotron sources. We introduce a novel image processing technique combining an eigenflat optimization with deformable image registration to address the challenges and enable quantitative high-resolution scans of centimeter-sized objects with multiple-micrometer resolution. We demonstrate the potential of the method by obtaining an electron density map of a rat brain sample 15 mm in diameter despite the limited horizontal field of view of 6 mm of the beamline. This showcases the technique's ability to significantly widen the range of applications of modulation-based techniques in both biological and materials science research.

Online Dose Verification in VHEE Radiotherapy Using Bremsstrahlung Radiation

2026-03-18T15:21:36Z

Very high energy electrons (VHEE) in the 50-250 MeV range, delivered in short pulses at ultra-high dose rates, are proposed for clinical FLASH radiotherapy (RT) targeting deep-seated tumors. The clinical implementation of VHEE-FLASH RT requires online verification to optimize dose delivery. In this study we propose a novel online dose verification technique based on the detection of bremsstrahlung photons during VHEE interactions with matter. A polymethyl methacrylate (PMMA) phantom was simulated to evaluate the dose deposited by a VHEE beam and to optimise the system design to detect the bremsstralung radiation. Experimental validation was performed at the Beam Test Facility at Laboratori Nazionali di Frascati (Isituto Nazionale di Fisica Nucleare-INFN). A deep learning pipeline was developed to reconstruct the dose distribution in the phantom based on the bremsstrahlung radiation profile. Experimental results demonstrated successful detection of bremsstrahlung radiation emitted orthogonally to the beam axis. The deep learning model achieved accurate dose reconstruction based on the bremsstrahlung radiation profile with a discrepancy of less than 2 percent compared to the simulated dose distribution. This study confirms that bremsstrahlung detection provides a viable online verification for VHEE RT.

The FLASH enigma

2026-03-18T13:05:28Z

We consider physics behind the FLASH modality of cancer radiation treatment where extremely short treatment times are achieved with ultra high dose rates maintaining the conventional antitumor effectiveness and yet substantially reducing damage to normal tissues (sparing effect). The difference in responses between normal and tumor tissues is attributed here to different recombination rates related to their structure morphologies: ordered in normal vs disordered in the tumor tissues. Correspondingly different are their charge densities under ionizing radiation. In normal tissues it is high enough to form electron-hole liquid (EHL). Because of low EHL diffusivities, the chemical reaction and generation of free radicals are suppressed; hence, sparing effect. To the contrary, a disordered tumor tissue renders efficient energy relaxation channels forming antitumor free radicals. We describe the FLASH thresholds for doses and dose rates.

Distributional Consistency Loss: Beyond Pointwise Data Terms in Inverse Problems

2026-03-17T13:39:21Z

Recovering true signals from noisy measurements is a central challenge in inverse problems spanning medical imaging, geophysics, and signal processing. Current methods balance prior signal priors (regularization) with agreement with noisy data (data-fidelity). Conventional data-fidelity loss functions, such as mean-squared error (MSE) or negative log-likelihood, seek pointwise agreement with noisy measurements, often leading to overfitting to noise. In this work, we instead evaluate data-fidelity collectively by testing whether the observed measurements are statistically consistent with the noise distributions implied by the current estimate. We introduce distributional consistency (DC) loss, a data-fidelity objective that replaces pointwise matching with distribution-level calibration. DC loss acts as a direct and practical plug-in replacement for standard data consistency terms: i) it is compatible with modern unsupervised regularizers that operate without paired measurement-ground-truth data, ii) it is optimized in the same way as traditional losses, and iii) it avoids overfitting to measurement noise without early stopping or priors. Its scope naturally fits many practical inverse problems where the measurement-noise distribution is known and where the measured dataset consists of many independent noisy values. We demonstrate efficacy in two key example application areas: i) in image denoising with deep image prior, using DC instead of MSE loss removes the need for early stopping and achieves higher PSNR; ii) in medical image reconstruction from Poisson-noisy data, DC loss reduces artifacts in highly-iterated reconstructions and enhances the efficacy of hand-crafted regularization. These results position DC loss as a statistically grounded, performance-enhancing alternative to conventional fidelity losses for an important class of unsupervised noise-dominated inverse problems.

Fractal and Spectral Dimensions as Determinants of Thermal Ablation Outcomes in Cancer Tissues

2026-03-17T13:26:27Z

Clinical thermal ablation outcomes display significant variability that classical bio-heat models cannot fully explain. One reason may lie in the fractal architecture of biological tissues, which has been identified as a robust biomarker directly correlated with cancer grades. This structural heterogeneity, together with memory effects (e.g., thermotolerance), causes heat transfer in living tissues to differ from Fourier diffusion, resulting in anomalous biological transport. In this work, we implemented a realistic fractal-fractional bio-heat model, with non-linear perfusion and PI-controlled power delivery, to quantify the role of tissue fractality in ablation outcomes. Our results reveal that the expansion of coagulation zones is jointly controlled by fractal geometry and its associated topological connectivity. These findings highlight spectral dimension as a key driver of clinical variability, successfully reproducing the reduced ablative efficacy in liver metastases compared to primary carcinomas, and provide evidence for topologically informed treatment strategies for the thermal ablation of malignant neoplasms.

Axonal microstructure and compartmentalization impact the orientation and time dependence of mesoscopic transverse relaxation

2026-03-17T07:38:57Z

In biological tissue, MR transverse relaxation stems from mechanisms spanning multiple scales, from molecular dipole-dipole interactions to mesoscopic field variations driven by tissue microstructure. While mesoscopic relaxation reflects cellular organization, its dynamics in white matter, specifically its dependence on axonal orientation and echo time, remain less investigated. This study employs theoretical frameworks and Monte-Carlo simulations using 3D EM-based white matter (WM) geometries to investigate how compartmentalization and realistic morphology drive these effects. Specifically, we simulate intra-axonal relaxation driven by magnetic fields induced by realistic axonal myelin sheaths and intra-axonal spheres as a model of iron-containing mitochondria. Our results confirm that orientation dependence of mesoscopic relaxation in WM is detectable and agrees with experimental observations. The time-dependence aligns with 1-dimensional short-range structural disorder, but at clinical echo times, this signature may be masked by dominant molecular relaxation. This work moves beyond idealized models to aid the development of more specific biophysical models of mesoscopic relaxation to achieve better biomarkers for neurodegenerative disease.

Notes on the primal-dual algorithm for convex optimization applied to X-ray tomographic image reconstruction

2026-03-16T16:59:07Z

The purpose of these notes is to provide background on understanding the primal-dual algorithm of Chambolle and Pock [1] for imaging scientists. The presentation focuses on providing intuition and an algorithmic system that is amenable to pre-conditioning. The document aims to be self-contained, providing background on the essential facts of non-smooth convex analysis.[2]

Hamiltonian dynamics for stochastic reconstruction in emission tomography

2026-03-16T10:25:08Z

The AMIAS/RISE framework formulates emission tomography as a probabilistic inverse problem in which reconstructed images are sampled from a distribution defined by the measurement model and counting statistics. In this work we present a stochastic reformulation of this approach based on gradient-driven optimization combined with Hamiltonian Monte Carlo (HMC) sampling directly in high-dimensional voxel space. This formulation enables practical ensemble generation for tomographic image reconstructions and provides direct access to image fluctuations within the sampled ensemble. Beyond point reconstruction, we introduce a spatially resolved operator-weighted diagnostic-the sampled data-visible variance-which quantifies how image fluctuations propagate through the imaging operator and thereby probes the local conditioning of the inverse problem under realistic acquisition physics. Using controlled software phantoms, experimental anthropomorphic phantom measurements, and a clinical DATSCAN SPECT acquisition, we demonstrate that while point-estimate accuracy under ideal conditions is comparable to deterministic reconstruction methods, the stochastic formulation provides additional physically interpretable insight. In particular, the ensemble analysis distinguishes uncertainty arising from intrinsic ill-posedness of the inverse problem from that associated with forward-model inadequacy. The clinical example is included to illustrate methodological applicability under realistic acquisition statistics rather than to assess diagnostic performance. The results establish the stochastic reconstruction framework as a practical ensemble-based approach for uncertainty quantification and forward-model validation in emission tomography.

A hybrid statistical sampling and iterative regularization method in sparse-view computed tomography

2026-03-16T06:08:11Z

Sparse-view computed tomography (CT) is an effective method to reduce the radiation exposure in medical imaging. To reduce the severe streaking artifacts that occur in reconstructed images due to violation of the Nyquist/Shannon sampling criterion, regularization is widely used to minimize the cost function. However, the iterative methods may lead to the accumulation and propagation of errors, which adversely affects the restoration of image details and textures. In this paper, we propose a hybrid model that integrates statistical sampling with iterative regularization to simultaneously shorten the sampling time and enhance the reconstruction quality. The proposed method is validated using three datasets: the Shepp-Logan phantom, the actual walnut X-ray projections provided by the Finnish Inverse Problems Society, and the clinical lung CT images.

A mathematical model of tumor growth using fractional derivatives

2026-03-15T19:28:36Z

In this work, we investigate a fractional-order tumor growth model aimed at capturing memory effects and nonlocal temporal dynamics inherent to tumor evolution. The model is formulated using Caputo fractional derivatives and incorporates key biological mechanisms related to tumor growth, vascular interaction, and cell clearance. To numerically solve the resulting fractional differential equations, a second-order fractional Runge-Kutta scheme is derived based on a truncated fractional Taylor expansion, providing an accurate and stable computational framework. The proposed model is calibrated using experimental tumor volume data from five patients, and its performance is evaluated through the Root Mean Square Deviation (RMSD) between numerical simulations and experimental observations. The results show that, for all patients considered, the fractional-order model significantly improves the agreement with experimental data compared to the classical integer-order formulation. An optimal fractional order α < 1 is identified in each case, highlighting the relevance of memory effects in tumor growth dynamics and their patient-specific nature. Further insight is obtained through phase-space and projection analyses, which reveal substantial geometric differences between integer-order and fractional-order dynamics. Although the present study is based on a limited number of patient datasets, the results demonstrate the potential of fractional-order modeling as a flexible and powerful framework for describing individualized tumor growth behavior. The proposed approach provides a solid basis for future extensions involving larger datasets, uncertainty quantification, and the incorporation of treatment effects and control strategies.

Retrieval of multiple fibre orientations using X-ray dark-field signal modelling

2026-03-15T19:04:02Z

Dark-field imaging is widely used to infer fibre orientation from signal modulation as a function of sample orientation. However, current X-ray dark-field retrieval methods are restricted to single orientations and require tomography to resolve overlapping structures. This approach is time-consuming and not suitable for thin materials, which are common in materials science. Here we present a dark-field model capable of retrieving multiple fibre orientations within a single pixel. The model, based on a geometrical description of fibre scattering, was validated through Monte Carlo simulations and experiments using beam-tracking setups with 1D and 2D masks. Results demonstrate reliable orientation retrieval for up to two fibres per pixel, with the 2D mask providing multi-directional sensitivity in a single acquisition and enabling faster and simplified data collection.

First in-vivo human magnetic particle imaging

2026-03-14T09:12:33Z

Magnetic particle imaging (MPI) is a tracer-based technique that directly detects the distribution of magnetic iron-oxide nanoparticles with millisecond temporal resolution and no tissue background. Despite extensive preclinical work, in-vivo application of MPI in humans has not previously been reported. Here, we report the first in-vivo human MPI angiography, visualizing venous perfusion of the upper extremity using a human-scale scanner and clinically approved ferucarbotran. Under identical procedural conditions, we performed X-ray digital subtraction angiography as the clinical gold standard. MPI visualized major superficial and deep veins, including inflow, branching, valve filling, and clearance dynamics in real time with 2 frames per second. These results establish magnetic particle imaging as a clinically translatable modality for radiation-free vascular imaging in humans and mark the transition of MPI from preclinical research to first clinical application.

Multi-contrast laser endoscopy for in vivo gastrointestinal imaging

2026-03-14T00:36:00Z

White light endoscopy is the clinical gold standard for detecting diseases in the gastrointestinal tract. Most applications involve identifying visual abnormalities in tissue color, texture, and shape. Unfortunately, the contrast of these features is often subtle, causing many clinically relevant cases to go undetected. To overcome this challenge, we introduce Multi-contrast Laser Endoscopy (MLE): a platform for widefield clinical imaging with rapidly tunable spectral, coherent, and directional illumination. We demonstrate three capabilities of MLE: enhancing tissue chromophore contrast with multispectral diffuse reflectance, quantifying blood flow using laser speckle contrast imaging, and characterizing mucosal topography using photometric stereo. We validate MLE with benchtop models, then demonstrate MLE in vivo during clinical colonoscopies. MLE images from 31 polyps demonstrate an approximate three-fold improvement in contrast and a five-fold improvement in color difference compared to white light and narrow band imaging. With the ability to reveal multiple complementary types of tissue contrast while seamlessly integrating into the clinical environment, MLE shows promise as an investigative tool to improve gastrointestinal imaging.